Big Data in Healthcare: Using Analytics to Identify and Manage High-Risk and High-Cost Patients

The adoption of electronic health records in healthcare is higher than ever (around 89%), which significantly increases the quality of clinical data available for analysis. Consequently, rapid progress has been made in clinical predictive analytics techniques. The market of Big Data solutions in healthcare was valued at USD $23,749.33 million in 2020 and is forecasted to grow to USD $58,404.24 million by 2026.

Predictive big data analytics for healthcare helps to identify and manage high-risk and high-cost patients, ensuring better care outcomes and reducing healthcare costs. In this article, we will discuss the main reasons to use analytics and Big Data for healthcare, especially when working with these high-risk, high-cost patients.

High-Risk and High-Cost Patients in Healthcare

Total national healthcare spending in the US reached $3.8 trillion in 2019. Meanwhile, approximately 5% of patients require 50% of health care system funding. Reducing the expenses for these patients and optimizing their treatment plans is of utmost importance for healthcare providers. These 5% of patients are considered high-cost because of their high-risk conditions.

High-risk patients are defined as those who suffer from one or more chronic illnesses, have unstable conditions, and require constant, costly treatment. Compared to low-risk groups, high-risk patients require a lot of effort, attention, and time from medical personnel. Experts divide high-risk patients into 3 cohorts:

Patients with a One-Time Catastrophic Health Event

The first cohort consists of people who experienced a single catastrophic health event, such as major trauma or a sudden life-threatening illness like acute cancer. The condition of these patients is hard to predict, but they have a chance to recover.

Patients Suffering from Chronic Illness and Whose Condition Can be Improved

The second cohort consists of patients who suffer from one or multiple chronic conditions such as stable heart failure, diabetes, asthma, or mental illnesses. They often visit emergency rooms and doctors and require periodic hospitalization. The main focus of disease-management programs for this cohort is to improve patient health and reduce medical costs.

Patients Suffering from Chronic Illness and Whose Condition Cannot be Improved

The third cohort consists of people suffering from several severe medical conditions such as heart failure and chronic renal disease. These are people from the second cohort whose condition deteriorated over time and their health can’t be restored now. Patients in this cohort require costly medications, therapies, and intense interventions till the end of their lives. Their treatment is adjusted to prevent acute exacerbation.

Source: hcinnovationgroup.com

The treatment approaches differ for these 3 cohorts, but utilizing the power of predictive analytics helps hospitals manage all types of high-risk patients. Analyzing the patients’ unique health patterns and personalizing their treatments results in improving the quality of care and reducing medical expenses.

How to Identify High-Risk Patients

The identification process of high-risk patients includes two major steps: data gathering and interpreting. The more data one has, the more accurate the results will be.

These are the metrics that are being taken into consideration while using predictive analytics in healthcare for identifying high-risk patients:

• age
• gender
• number of emergencies per year
• laboratory test results
• examination results from specialists
• number of hospital admissions per year
• chronic illnesses
• drugs taken
• hospital visits
• rates of not showing up for appointments
• mental health condition
• self-reported health statuses
• This data can be found in electronic health records, administrative claims, health risk assessments, and clinical input.

However, non-medical data plays a huge role in the identification as well. It includes lifestyle, environment, socioeconomic condition, health literacy, view on healthcare (some patients might not believe in medicine), functional and cognitive impairments, etc.

Predictive systems analyze all this data to predict the severity of patients’ condition, the possibility of positive treatment outcomes, and the chances of patients turning out to be high cost.

A good example of using the power of predictive analytics in healthcare is the Emergency Department of NorthShore University HealthSystem. NorthShore University HealthSystem, which operates four hospitals in Illinois, developed a tool that physicians and nurses in the emergency room can use to identify patients at high risk of a heart attack.

The tool assigns each patient a score between 1 and 10, indicating how likely they are to be at risk of a heart attack. Using the tool helps doctors and nurses decide which patients don’t need observation, allowing those patients to be sent home. This solved the problem of long wait times, lack of beds, and wasted time for medical personnel.

Benefits of Using Predictive Analytics and Big Data in Healthcare

Big Data and healthcare analytics make it possible to analyze large amounts of data at once. These analytics will not only take into account a current patient’s data but will also compare it to previous similar cases.

Why do so many healthcare institutions rely on Big Data and try to implement predictive analytics in healthcare? Because of the immense benefits, including the following:

Adjusting Care Plans for Better Outcomes

One of the most difficult aspects of providing healthcare to high-risk patients is creating a personalized approach. Analyzing the previous medical history of a high-risk patient with Big Data and comparing it with similar cases helps the hospital to adjust the treatment plan accordingly.

Optum Lab, the US research and innovation center, has created a database of EHRs of over 30 million patients that can be used by predictive analytics tools to enhance the quality of care provided. Their main goal is to enable healthcare professionals to make well-informed, data-driven decisions that reduce healthcare costs.

Predicting Changes in Patient’s Condition

46% of patients suffer from OIRD (Opioid-Induced Respiratory Depression), which causes over 50% of medication-related deaths in healthcare organizations—97% of which could have been prevented.

Using predictive clinical analytics tools to evaluate continuous data streams from multiple patient monitors, along with retrospective information from EHRs, provides a holistic picture of a patient’s health condition. Constant surveillance is especially relevant for high-risk patients as their condition is unstable. When the analytics tool detects potential deterioration, it notifies doctors, so they have time to adjust treatment plans before it happens.

Uncovering Unique Health Patterns

Each year in the US, over 250,000 patients experience adverse drug events that are not only expensive but also cause substantial morbidity and mortality. The use of Big Data analytics in healthcare helps hospitals better analyze the unique health patterns of a particular patient and prevent adverse drug events.

Foreseeing Potential Readmission

Under the Hospital Readmissions Reduction Program, Medicare can withhold up to 3% of regular reimbursement for healthcare institutions that have a higher-than-expected number of 30-day readmissions for the defined conditions. In 2020, 83% of hospitals were penalized for a total of $563 million because of the high amount of readmissions.

Predictive analytics solutions in healthcare can help healthcare institutions predict potential readmissions and better coordinate care after discharge for high-risk patients.

Predicting the Number of Patients

Many hospitals use Big Data analytics in healthcare to forecast the number of patients they will need to manage in a coming time period. This helps to better manage staffing, patient transfers, and available beds, which is especially relevant for the Emergency Department. A 10-minute increase in the waiting time for them on average increases the hospital’s cost for high-risk patients by 6%.

Source: accenture.com

Our Work

Langate healthcare software development company has been working in the niche for more than 10 years now. Our professional developers have created a variety of custom solutions for the industry, including data exchange tools and clinical trial management software.

Approved Admission, one of our clients, needed to develop a compliant insurance eligibility validation software that would facilitate data exchange between different IT platforms. Langate created a robust infrastructure that retrieves, processes, and analyzes information from Medicare, Medicaid, and HMO insurance payers.

Because it provides a flexible data exchange, Approved Admission has managed to become a market leader that increases its user base each year.

Bottom Line

Identifying high-risk patients is essential in healthcare and may save a lot of money later. However, detecting them is a complex process.

Many factors should be analyzed before deciding on a patient’s status, including both medical (chronic illnesses, hospital visits, mental health condition, laboratory test, specialist examination, etc.) and non-medical (lifestyle, environment, health literacy, functional and cognitive impairments, etc.) conditions.

Big Data in high-risk and high-cost patient identification helps to make more accurate decisions, avoid clinical mistakes, improve medical outcomes, and reduce healthcare costs.

If you’ve realized that you are in need of deploying predictive analytics in healthcare using Big Data, do not hesitate to reach out to Langate. Our experienced developers and project managers will provide high-quality Big Data development services for your high-risk patient identification and management.